Discovering the driving factors of landslides of the Wenchuan earthquake using remote sensing and GIS

On May 12 in 2008, a Magnitude 8.0 earthquake hit Wenchuan in China, and the casualty shocked the whole world. The landslide was a frequent secondary disaster in this earthquake, so to analyze the mechanism of landslides in the disaster area is very important for post-earthquake reconstruction. The study area is located in PingWu County, which was also hit by the earthquake severely. And the data sources are ETM+ image, DEM and interpreted ALOS image. This paper considered four potential driving factors for landslides, and they are land cover, lineament, slope and drainage. The land cover was classified based on the density of vegetation, and sub-pixel analysis was employed; Density of lineament was calculated by Sobel operator and image segmentation; Slope was classified by using a threshold; Drainage was considered without numerical analysis, because it is significant and simple in study area. To find out how they influenced the landslides, conditional probability was utilized as a measurement. The result shows that areas in sparse vegetation, dense lineament and steep topography were easy to meet landslides, while drainages also induced landslides. ? 2008 SPIE.EI